300 THE DEVELOPMENT OF THE VENTRAL NERVES IN SELAGHII. 



Evidence that they are the nuclei of the cells which form the neurilemma sheath will 

 be given later. 



Third, in the adult, as in the early stages of development, the neuraxones whose 

 cellular connections have been traced are found to be processes of neuroblasts (gang- 

 lion cells) in the ventral horn of the neural tube. This fact and the fact that no gang- 

 lion cells are found in the ventral nerves of the embryo or the adult cannot be recon- 

 ciled with the assumption that neuroblasts migrate into the nerves, unless the further 

 assumptions are made to the effect that the neuroblasts become metamorphosed 

 into neurilemma cells and that the proximal processes of such cells effect a secondary 

 union with ganglion cells in the ventral horn of the spinal cord. I have looked in 

 vain for evidence to support either of these assumptions. The conclusion seems 

 unavoidable that the migration of the neuroblasts is confined within the wall of the 

 neural tube. The cause of migration appears to be the multiplication of cells in the 

 inner layer of the medullary walls. 



The neuroblasts take a ventro-lateral position in the wall of the neural tube just 

 within the marginal veil of fibres forming the white matter of the cord, where they 

 soon become multipolar by the development of dendritic processes. The latter first 

 appear after migration has ceased. His ('89) regards the formation of the marginal 

 veil and of the neuroglia network as the mechanical cause for the cessation of neuro- 

 blast migration. The two phenomena appear correlated in Squalus, but it is difficult 

 if not quite impossible to say that the relation is a causal one. 



Four chief stages in the differentiation of the nidulus of a ventral nerve may then 

 be distinguished: (1) the stage of the undifferentiated columnar epithelium; (2) the 

 formation of the neuraxon process;* (3) the lateral migration of the pear-shaped 

 unipolar neuroblast; and (4) the formation of dendritic processes. 



2. The Histogenesis of the Neuraxon and the Neurilemma A. Historical. 



Although many subordinate problems are involved in the general one of the histo- 

 genesis of nerve fibres, only the three following are discussed in this paper: 



a. Is the neuraxon of a ventral nerve fibre multicellular in origin or the process 

 of a single medullary cell? 



b. Are the cells which form the neurilemma emigrated medullary elements (ecto- 

 dermal) or mesenchymatous (mesodermal) cells? 



c. Are the ventral nerves db initio connected with their terminal organs and 

 how intimate is this connection? 



None of these three questions, about which the greatest morphological interest 

 centres, has been finally answered for Selachii. 



* It is probable that for some neuroblasts the second and third stages are reversed. The proportion of such 

 neuroblasts to the whole number of neuroblasts has not been determined. 



